A new technique for performance evaluation of random access protocols

A new technique is proposed for performance evaluation of random access protocols. This technique is based on the idea that the service probability distribution of a particular random access protocol can be described by a phase-type distribution. By modeling the service probability distribution of a protocol into a phase-type distribution, the performance of the protocol can be evaluated using an equivalent continuous time Markov chain. We demonstrate the applicability of the technique to evaluate the performance of the IEEE 802.3 and the IEEE 802.11 MAC protocols. Furthermore, we provide an example of performance evaluation of a MAC protocol subject to bursty traffic models such as the Markov modulated Poisson process. The accuracy of the results is verified by simulation.